ABSTRACT Trabecular bone (TB) is a complex quasi-random network of interconnected struts and plates. TB constantly remodels to adapt dynamically to the stresses to which it is subjected (Wolff's Law). In osteoporosis, this dynamic equilibrium between bone formation and resorption is perturbed, leading to bone loss and structural deterioration, both increasing fracture risk. Most osteoporotic fractures occur at sites rich in TB (vertebrae, radius, proximal femur). Bone's mechanical competence can only be partly explained by variations in bone mineral density (BMD), which led to the notion of bone quality, chief among which is architecture as a determinant of TB strength. Recent advances in CT and MRI now allow imaging of TB in vivo. However, the limited SNR precludes voxel sizes much smaller than TB thickness (100-150m), therefore resulting in images that are inherently fuzzy. Therefore, most conventional histomorphometric approaches to TB structure assessment are not applicable to in vivo resolution regime. This proposal introduces a new morphometric index called tensor scale (t-scale) to measure quality of TB micro architecture via in vivo imaging and designs experimental plans to evaluate reproducibility and sensitivity of t-scale-based TB architectural measures. Also, we will examine their strengths to detect TB architectural changes in response to disease or treatment progression. The fundamental principle of t-scale is to fit an ellipsoid to a local structure. The unique property of t-scale is that the ellipsoid's shape, orientation and size simultaneously determine the topology (plates vs rods), orientation and scale of trabeculae. Our major goals in this project are - (1) to develop the methodology for computing t-scale-based architectural measures from TB images, (2) to evaluate the sensitivity and reproducibility of t-scale-based measures, (3) to examine t-scale measures' ability to predict experimental biomechanical parameters of TB specimens and (4) to examine the sensitivity of t-scale measures to detect the effects of osteoporotic TB loss and antiresorptive treatment via in vivo MRI. The proposed method will (1) obviate the need for binarization, (2) characterize topology, orientation and scale without the need for skeletonization and (3) detect early TB architectural changes in response to treatment or disease progression. The central hypothesis is that the new parameters are more sensitive to detecting remodeling effects and more reproducible than conventional measures. Sensitivity and reproducibility of the new method will be evaluated using synthetic TB networks, micro-CT and ex vivo MR imaging of TB cores from cadaveric distal radii along with experimental biomechanical data, MR images of intact specimens and of human subjects, and finally, patient data from clinical studies previously funded by the NIH. Our objective is to apply t-scale based analysis methods to longitudinal and cross-sectional imaging studies for assessing bone quality.